The preparation and reactions of stable phosphorus ylides derived from maleic anhydrides, maleimides or isomaleimides

The preparation and reactions of stable phosphorus ylides derived from maleic anhydrides, maleimides or isomaleimides

Ttmhedrm. Vol. 24, pp. 2241 to 2254. Pergmmn Ras 1968. Printed in Great Britm THE PREPARATION AND REACTIONS OF STABLE PHOSPHORUS YLIDES DERIVED F...

869KB Sizes 29 Downloads 88 Views

Ttmhedrm.

Vol. 24, pp. 2241 to 2254.

Pergmmn Ras

1968.

Printed in Great Britm

THE PREPARATION AND REACTIONS OF STABLE PHOSPHORUS YLIDES DERIVED FROM MALEIC ANHYDRIDES, MALEIMIDES OR ISOMALEIMIDES E. HEDAYAand S.

THEODOROPULOS

Union Carbide Research Institute, P.O. Box 278, Tarrytown, N.Y. (Received in USA 19 July 1967; acceptedfor

publication

1 September 1967)

Ahatraet-A number of stable triphenylphosphoranylidenesuccinimide derivatives have been prepared from maleimides or isomaleimides and triphenylphosphine. Also a stable ylide has been obtained from citraconic anhydride and triphenylphosphine. The succinimide ylides were found to react with a wide variety of aldehydes giving the corresponding ylidene derivatives in high yields. Although these triphenylphosphoranylidenes did not react with ketones. a tri-n-butylphosphoranylidene derivative did react with cyclohexanone to give a cyclohexylidenesuccinimide. In contrast, the recently discussed triphenylphosphoranylidenesuccinic anhydride reacts with cinnamaldehyde or phenylpropopargyl aldehyde to give the corresponding diylidene succinic anhydrides (fulgides). Intractable mixtures were obtained with the other aldehydes investigated. Furthermore, the stable ylide derived from triphenylphosphine and N-phenylmaleimide gives a ketenimine with phenyl or methyl isocyanate or phenyl isothiocyanate. The resulting ketenimines react with amines or benzenethiol to give diaminoethylene derivatives (ketenaminals) or a phenylthioaminoethylene derivative, respectively. Also, this phosphorus ylide reacts with nitrosobenzene to give 3-anilino-N-phenylmaleimide. The synthetic utility of these stable phosphorus ylides is discussed as well as other features of their reactions. INTRODUCTION IN THE course of an investigation’ concerned with the chemistry of isomaleimides and maleimides we found that triphenylphosphine reacts with both of these isomers to give ylide adducts (1 and 2) having structures analogous to the recently discussed maleic anhydride adduct la.’ Subsequently, we discovered that, except for la, these readily obtained, stable adducts undergo the Wittig reaction with CO compounds to give derivatives which correspond to those which, in part, one would otherwise obtain from the Stobbe condensation.3 Because of the ease of preparation of these ylides and the high yields of Wittig reaction products obtained in our initial experiments with aldehydes. we embarked on an investigation to define their synthetic utility.

la: R = H, X = 0 b: R=H,X= NH

e: d: e: f:

3a: X=NPh b: X=NH

R=H,X=NPh R = H. X = N-NC4H,0 R=Me,X=O R = Me. X = NPh 2241

2242

E. HEDAYAand S. TH~~OLXX~OPUL~~ RESULTS

AND

DISCUSSION

Prepmotion ofylides 1, 2, and 3. The ylides 1 and 2 were prepared by combining an excess of triphenylphosphine and the maleic anhydride or maleimide in glacial acetic acid. In some cases the reaction mixture was warmed on the steam bath for periods of about f hr. The ylide was precipitated by addition of excess ether and recrystallized in most cases from acetone. The yields were generally 50% or better. Some spectral properties of these ylides are summarized in Table 1. TABLE1. SPECTRAL Anhydridc or Imide

Ylide

Maleicd Malemide N-Phenylmalemide N,N’-Morpholine maleimide Citraconic N-Phenylcitraconimide N.N’-Bimaleimide

la lb lc Id

le If

2

PROPERTIES

OF YLIDES

1 AND

2 Area Ratiob

G(PPmF

2.88d

34Om 0.82d @9Od -

3.25sb

7.6m

2:15

-

-

-

3.2Od 3.75m 3.4q 3.3m -

7.6m 7.5m 7.5m 7.6m

2:20 2:8:15 1:3:15 3:1:20

5.60 581 5.85 5.82 5.59 5.83 5.75

5.93’ 609 6.09 612 590 6-08 6a

a NMR obtained in deuterochloroform with tetramethylsilane as internal standard; s. singlet; sb. broad singlet ; etc. .b Nearest integral area ratios are listed in order of high field resonance to low tield resonance as in column 3. ’ Infrared spectra were obtained in Nujol mulls. ’ litzb 6 3.25s. 7.6m in deuterochloroform. ’ lit.zc*dI C=O 5.63. 5.94 p in Nujol mull.

The structure proof for the previously prepared triphenylphosphoranylidenesuccinic anhydride (la) was based on the IRzd and NMR spectra,20*b synthesis from chlorosuccinic anhydride,*“, b and methanolysis to a half acid which when treated with diazomethane gave authentic dimethyl triphenylphosphoranylidenesuccinate.*’ The correspondence of the IR CO bands between the adducts derived from maleimides and la along with NMR data (Table 1) indicate an analogous ylide structure for the former. Also, in our case chemical proof of structure for the triphenylphosphine adducts can be obtained from their characteristic Wittig reaction products (see below). Previous workers* either used benzene or acetone as a reaction solvent.for preparation of triphenylphosphoranylidenesuccinic anhydride (la). In our work we generally found that use of acetic acid as a reaction solvent led to more easily isolated, cleaner products. For example, Osuch et ~1.‘~were unable to prepare the citraconic anhydride-triphenylphosphine adduct (le) in ether, while in acetic acid we found that 1Cwas readily obtained as a stable crystalline solid. In contrast, the tri-n-butylphosphine adducts 3. which were similarly prepared in acetic acid, were found to be oils which were best characterized by the reaction with aldehydes (see below). Interestingly, when the isomaleimides 4l and 54 were combined with triphenylphosphine in acetic acid, the adducts 2 and lc were also isolated in high yields. It is not known whether prior rearrangement of the isoimide occurs, induced by triphenylphosphine. or if an initially formed triphenylphosphine-isoimide adduct rearranges to the imide isomer under the reaction conditions. Many examples of catalyzed and non-catalyzed 1.3 acyl 0 to N rearrangements of isoimides to imides

The preparation

and reactions

of stable phosphorus

ylides

2243

have been observed in both cyclic and acyclic systems.lb* ’ Actually, the triphenylphosphine adduct Id was obtained from the readily prepared isoimide 6 rather than the imide and the use of the biisoimide 4 for preparation of 2 is similarly more convenient than the use of N.N’-bimaleimide since the former is more readily obtained than the latter.l”

4

5

6

Reaction of ylides 1,2 and 3 tith aldehydes and ketones. It was previously reported that the vigorous reaction of benzaldehyde, pnitrobenzaldehyde, or butyraldehydeZb with triphenylphosphoranylidenesuccinic anhydride (la) gives an intractable product from which only traces of the Wittig product could be isolated. A better yield of dichlorovinylmaleic anhydride (25 p/,) was obtained from the reaction of chloral and la and this was interpreted as involving initial formation of the expected Wittig product. 2b We have confirmed the result with benzaldehyde and la. In contrast. the reaction ofa wide variety ofaldehydes with our triphenylphosphoranylidenesuccinimide derivatives proceeds smoothly giving high yields of the crystalline

Rc%wh ‘-tYH RC

78: R=Ph b: R = CH:CHPh c: R = pC,H,N

&: R=Ph b: R = pC,H,N c: R=Et

98: R=Ph b: R = CH:CHPh

RC

10: R = p-C,H,N

E.

2244

HEDAYA

and S. THEODOROPULOS

Wittig products. For example, lc when combined with a variety of aldehydes gives the anticipated Wittig products 7 in 70% yields or better. Similar results were obtained with lb, ld, and 2 giving the Wittig products &lo. and 9 respectively although a much more limited range of aldehydes were investigated. Generally, these reactions were carried out simply by mixing the aldehyde and ylide, warming the mixtures for a few minutes. and then letting them stand at room temperature for periods of l-5 hr. In some cases. such as 7c, 7g, 7b and 8b it was advantageous to use methanol or dimethylsulfoxide as a solvent. Purification was readily achieved by washing the crude reaction products with methanol or ether to remove triphenylphosphine oxide and recrystallization from a suitable solvent. An interesting example of reaction of a disguised dialdehyde is that between 2,5dimethoxytetrahydrofuran and the triphenylphosphoranylidene derivative lc in acetic acid which gives the biimide derivative 11 in 50 % yield.

Preliminary attempts to obtain Wittig reaction products from benzaldehyde and the methyl substituted ylides le and If were unsuccessful. These ylides appear to be somewhat more stable than the unsubstituted examples. Furthermore, we found that all of the triphenylphosphine ylides synthesized here did not react with ketones such as cyclohexanone, benzophenone. or acetone, which is consistent with the lack of reactivity of other resonance stabilized ylides with ketones6 In fact, acetone was often used as the recrystallizing solvent. Attempts to catalyze the reaction with ketones with benzoic acid’ or other organic acids were unsuccessful. However, the tri-nbutylphosphine derivative 31 did react with cyclohexanone giving the cyclohexylidene derivative 12 in 20 p/, yield. Also. the ylide 3b was most conveniently used to prepare the propylidine derivative 8c.

3a

+ A

8

“1” N-Ph

12

The structure proof for all these Wittig reaction products is based on the elemental composition. the mode of synthesis, and the NMR and IR spectral data. Actually, our structural problem involves the distinction between the alkylidene or arylidene succinimide structure (13 or 14) and the alkyl or aryl maleimide structure (15) as well as the problem of geometrical isomerism exemplified by structures 13 or 14. With respect to the former the NMR spectra (Table 2) of the alkylidene derivatives 7d.

2.2Om

1.62sb

765s 2.70m 308m

3.92d 7.95s 2+3,lm 4.15d

4.08s 1.151

1.8-2.2m

3.75sb

7.5m 4649m

3.88d 3.59m

3.33sb

802s 0.8-3.lm

7.5m

418d

3.95d

9.1 t(T) 7,5m(T) 8.98d(T) 6&76m(C) 8.12s(C + T) 6.9m-7.45m(C) 7.2-7,5m(C) 7.831(T) 8.95d(T) 6.8tt(C) S.lOt(T) 7.0-8.Om(T) 89Od(C + T) 7.2277m(T) 7.4m(C)

7.8Od 63-6&h 3.5Od 7.52s 8.3Od 3.32sb 7.68s 3.92sb 8Gd 3.8Qsb 3.33s

PROPERTIES

7.52m 3.784 8.3Od 3.41d

d(ppmY

SPECTRAL

REACTION

3.2-3.8br

3.00

PRODUCTS

with structure.

C deuterochloroform;

2:5:1 2:1:2:2 3:2:2: 1 2:5:1 4:l 4:6:1:2:2 2~2~6 6:2:2:2:5

2:7:1:1 4:2:2:1:1:5 _d

2:lO:l 2:13 2:5:1x2:2 7:2:8

Area Ratio”

OF WIITIG

D NMR obtained in solvents indicated with tetramethylsilane as internal standard; singlet; etc. b (footnote’b-Table 1). c (footnote c-Table 1). d Too insoluble to obtain precise integral ; approximate area ratios were consistent

5 7b 8a 8b Rc A 9iJ 10 11 12

78 7h 7c 76 7e 7f

Product

TABLE 2.

6.02 630 6.05

6.02 5.80br 5.85 599 59Ob

sb, broad

6.09

605

6.10

6.10 610 6.05 6.00 5.95 606 5.95 6.06 605br 5.89

acid; s, singlet;

5.88 5.88 5.86 5.89 5.89 59Obr 5.85 590 5.8br 5.70 60br 5.75 5.68 5.61 5.89 5.68 T. trifluoroacetic

5.68

564 5.68 5.59 5.65 5.68 5.68 5.65 5.69 5.70 4Qbr 3.3br 5.62

E. HEDAYAand S. THE~D~ROPUUX

2246

7g, & 11. and 12 clearly rule out the maleimide structure 15 for these derivatives. In particular the resonance at 3.34.2 ppm for the succinimide ring methylene protons generally appear as an approximate doublet (J x l-2 c/s). a broad singlet, or for the case of 12 as a sharp singlet. Also, the vinyl proton resonance for & appears as a complex multiplet (approximately a triplet of triplets) which further rules out the maleimide isomer for this product. Since the tendency of the double bond to rearrange into the imide ring under the reaction conditions should decrease for the arylidene derivatives compared to the alkylidene derivatives and also since the succinimide methylene proton resonances in the former closely parallel those of the latter. all of these Wittig reaction products must have the general structure 13 or 14.*

13

14

15

With respect to the problem of geometrical isomerism. we have only limited experimental information at present. In all cases. crystalline, rather sharp melting products were obtained which suggests the predominance of one isomer. Also, the NMR spectra generally lacked any feature clearly suggesting the presence of geometrical isomers. Thus. it is probable that our reaction product consists predominantly of one isomer. It is generally accepted that resonance stabilized phosphoranes (e.g. carbomethoxymethylenetriphenylphosphorane) give predominantly one oletinic isomer in the Wittig reaction with aldehydes. which is the one with the alkyl or aryl group derived from the aldehyde being trans to the carbomethoxy group.’ The most pertinent example involves the reaction of methyl carbomethoxymethylenetriphenylphosphorane with acetaldehyde which gives a 96:4 ratio of 16 and 17 in 90% overall yield. lo This result was interpreted in terms of the relative stability of the two most probable betaine conformations. Analogously, the major product in our reactions is probably the isomer with the alkyl or aryl group derived from the aldehyde being tram to the imide carbonyl(l4).

/Me

W’=kOOMe

Me

+

Hpo

-

Me,

Pe

HRC=C \OO~e 16 (96 p/d)

Me

COOMe

‘ccc’

H'

‘Me 17(4 %I

Although the reaction of la with most aldehydes generally gives intractable mixtures. we have discovered that la does react cleanly with cinnamaldehyde to give a good yield of a red. relatively insoluble crystalline product which on the basis of the elemental analysis and IR spectrum has the fulgide structure 18. This was confirmed by synthesis of 18 by the Stobbe condensation of cinnamaldehyde with sodium succinate.” although in very low yields (0.5%). Consequently, the Wittig route is clearly a superior method for the synthesis of 18. We also found that phenylpropargyl aldehyde reacts with la to give the corresponding fulgide 19 in 55 y/, yield. * In a recent note the reaction of ethylmaleic ethyl maleimide and ethylidene isomers.’

anhydride

with ammonia

was shown

to give both the

The preparation ,a

and reactions of stable phosphorus

PhCH: CH

ylides

2247

CHO,

PhCH:CH*C

sodium succinate

C,PhCH:CH*C

PhCH:CH.CHO/

The above results suggest that one problem with the Wittig reaction of la may involve polymerization or telomerization of the initially formed ylidene derivative via secondary active methylene condensation reactions. Since the fulgides 18 or 19 are quite insoluble in the reaction mixture they are diverted from further reactions by precipitation. On the other hand. our success with ylides derived from maleimides may result from the reduced acidity of the succinimide ring methylene proton compared to corresponding protons of the succinic anhydride derivatives. Reaction of ylide lc dth isocyanates. isothiocyanates. and nitrosobenzene. The reaction of benhydrylidenetriphenylphosphorane with phenyl isocyanate, which gives triphenylketenimine. lz was one of the first Wittig reactions. This reaction has been more recently explored by Trippett and Walker.’ 3 They found that depending on the extent of substitution on the triphenylphosphinemethylene either acylation or ketenimine formation occurred as shown below. R’ R’, R+H

R’ \

C=P(Ph,)

R

>C=C=N-Ph

+ PhNCO

/ L

R=H

“‘.C=,Ph O=C

/

3)

\ NHPh

Consistently, we found that when the ylide lc was combined with either phenyl or methyl isocyanate or phenyl isothiocyanate. the corresponding ketenimines 2fl were obtained as crystalline solids in high yields. These derivatives had characteristic IR absorption at 49Ou and consistent NMR spectra. However, they are best characterized by their reaction products with amines. In contrast to triaryl ketenimines such as diphenylketene N-p-tolylimine where amidine derivates are obtained,r4 the ketenimines 2II gave excellent yields of the diaminoethylene derivatives (keteneaminals) 21. In most cases the crude ketenimine was used directly in these reactions. The proof of structure for 21 primarily involves the NMR spectral data along with the IR spectra and the elemental composition. The NhlR data (Table 3) in particular

2248

E. HFDAYAand S. THEODOROPU~

The

preparationand

2249

reactions of stable phosphorus ylides

unequivocally rule out the amidine structure 22 since a singlet was generally observed at 3.43.5 ppm corresponding to 2 succinimide methylene protons. A similar product (23) was obtained from 2Oa and thiophenol in 57 YYyield.

Amine -

2&a: R=Ph b: R=Mc

21r:R=Ph,R’=~ 3 b: R = Ph. R’ = NEt, c:R=Ph,R’=Na d: R=Me,R’=N 3

/

N-Ph

phN”c PhS’ c

23

These results are consistent with the inference that 21 has more stability than 22 because of the increased opportunity for conjugation in 21 compared to 22. On this basis. even if 22 were the initial reaction product, rearrangement to 21 would occur in the basic reaction medium. Similar considerations would apply to 23. We also found that lc reacts with nitrosobenzene, as does fluorenylidenetriphenylphosphorane.’ 5 to give the maleimide 24. The proof of structure is based on the NMR which reveals no resonance from succinimide methylene protons and one singlet resonance corresponding to a maleimide olefmic proton. The same product was recently reported i6 by Awad et al. to result from the thermal decomposition of the triaxoline. 25. Our IR spectra corresponds exactly with that given by these workersI for 24.

lc

+

PhNO

This result suggests that the imine tautomer of 24, which presumably is the initial reaction product. is less stable than 24. In contrast, the isoelectronic benzylmaleimide would appear to be less stable than N-phenyl benxylidenesuccinimide (7a) to the extent that the observed Wittig reaction result reflects thermodynamic stability. The question of the stability of 24 compared to its tautomer clearly involves a number

2250

E. HEDAYAand S. THEOD~ROPUL~~

of factors such as the relative C+N and C=C bond strengths. the relative strengths of NH and CH bonds in the two tautomers and of course the relative conjugation energies associated with the respective n systems of the two tautomers. Conclusion. The most significant features of these reactions are their simplicity and high yields. The preparation of the ylides as well as the Wittig reactions in all cases was most conveniently carried out in open Erlenmeyer flasks or even test-tubes and the products were all readily purified crystalline materials. The inertness of these ylides to ketones suggest the possibility of using them as a specific reagents for characterization of aldehydes. Alternatively. if reactions with ketones are desired. it is conceivable that this may be promoted by using ylides derived from trialkylphosphines. The alternate synthetic method for many of the compounds reported here would involve the Stobbe condensation as a key step. However. the Stobbe condensation with aldehydes generally gives relatively low yields of ylidene derivatives and a number of by-products because of the highly basic reaction medium.3 The use of potassium t-butoxide eliminates these complications to a large extent.3. ” However, the synthetic procedures involved in carrying out reactions with potassium butoxide are clearly less convenient than the Wittig reactions described here. Consequently, it is conceivable that the latter synthetic methods may find applications which formerly would involve the Stobbe condensation.3 Although an exhaustive survey was not carried out. it is shown that the ylides derived from maleimides will undergo reactions characteristic of resonance stabilized ylides other than reactions with aldehydes. These ylides should also react with imines. ketenes. la azides.’ 9 and epoxides.”

EXPERIMENTAL ~iphenylphosphorrmylidenesuccinic anhydride (la). MaJeic anhydride (29 g, 002 mole) and 6-O g triphenylphosphine were stirred in 30 ml glacial AcOH for 3 hr. Ether was added and the crystalline ppt WBS rccrystallizd from acetone giving 3.3 g (46”/‘J, np. 157-159” dec (lit.’ m.p. 1625-1635” dec). ~~phenylphoJphor~ym?ylidenesuccinimide (lb). Maleimide (10 g) and 2.6 g triphenylphoaphine were stirred in 20 ml glacial AcOH for 30 min on the steam bath. Work-up as above gave 3.6 g (- 100 y$, mp. 220”. (Found: C. 73.32; H, 527; N, 414. Calc for C,,H,sNO,P: C, 7352; H, 504; N, 3.89x.) N-Phenyl~iphenylphosphorMylidenesvccinim~e (1~). N-phenylmaleimide (09 g) and 1.5 g triphenylphosphine were stirred in 20 ml glacial AcOH on the steam bath for 1 hr. The reaction mixture was cooled and worked up as above giving 15 g (66 “/a m.p. 176%178.5”. (Found: C, 77.88; H, 5.12; N, 340. Calc. for C,sH,,N02P: C, 77.22; H, 5Q9; N, 3.21x.) The same product was obtained in 44% yield when N-phenylisomaleimide’ was substituted for Nphenylmaleimide in the above reaction. Methylniphenyylphosphoranylidenesuccinic anhydride (le). Citraconic anhydride (3.4 g, 003 mole) and 6.6 g triphenylphosphine in 30 ml glacial AcOH were stirred on the steam bath until a homogeneous soln was obtained (ca. 10 min) and then stirred 1 hr at room temp. Work-up as above gave 5.9 g (52 %), m.p. 179-181”. (Found: C, 73.68; H, 5.29. Calc. for C,,H,,O,P: C, 73.78; H, 511x.) Methyl-N-phenyyltriphenylphosphoranylidenesuccinimide (If). N-phenylcitraconimide (I.87 g, 01 mole) and 2.7 g triphenylphosphine in 3.0 ml glacial AcOH were stirred on a steam bath for 2 hr. Work-up as above gave 2.1 g (47x), m.p. 184-186”. (Found: C, 7723; H, 5.33; N, 292. Calc. for C2,H,,NOIP: C, 77.49; H, 5.38; N, 3.11%.) N,N’-Bitriphenylphosphoranylidenem&indde (2). N,N’-Biisomaleimide* (la g, 0005 mole) and excess triphenylphosphine in 20 ml Ac20 were stirred on the steam bath for 1 hr. Work-up as above gave the adduct in quantitative yield, m.p. ~350 Purilicatlon was achieved by trituration with ether. (Found: C, 73.26; H, 5.02; N, 4.02. C&z. for C2,HI,N0,P: C, 73.73; H, 478; N. 3.90x.)

The preparation and reactions of stable phosphorus ylides

225 I

The same product was obtained when NJ’-bimaleimidc’ was substituted for the isomakimide in the above reaction. N.N’-Morpholinorriphmylphosphoranylidenee (Id). N,N’-morpholinoisomaleimide (1.82 g. 001 mole), which was prepared from maleic anhydride and N-aminomorpholine as shown below, and 2.7 g (001 mole) triphenylphosphine in 20 ml glacial AcOH was stirred for 30 min on the steam bath. The soln WTISthen let stand for 30 min at room temp. Work-up as above 2.1 g (So%), m.p. 132-134”. This compound was further characterized by its reaction with pyridine&rboxaldehyde (see below). N-Morpholinoizomaleimide. N-aminomorpholine (5.1 g) and maleic anhydride (60 g) were stirred in glacial AcOH for 1 hr. After Ntration a nearly quantititive yield of the maleamic acid was obtained, m.p. 177-179”. This was cyclized to the isoimide in excess trifluoroacttic anhydride.‘* After removal of excess trifluoroacetic anhydride the residue was crystallized from ether to give 1.6 g of product, m.p. 121-123”. The IR had characteristic isoimide bands and the NMR was consistent with the structure. (Found: C, 52.76; H. 5.56; N. 15.47. Calc. for C,H,,N,O,: C, 52.73; H, 5.53; N, 15.38x.) Benzylidene-N-phenylsuccinimide (78). The ylide lc (la g, 00023 mole) was mixed with excess benzaldehyde in a tat-tube.. An exotherm was observed. After I5 min the mixture was triturated with MeOH and tiltered. The product (048 g, 79%) was recrystallized from benzene, m.p. 193-195”. (Found: C, 7760; H, 5.15; N. 5.56. Calc. for C,,H,,NO,: C, 77.57; H, 4.97; N, 5.32x.) CiMamolidene-N-phenylsuccinimide (7h). The ylide lc (10 g, 23 mmoles) and excess cinnamaldehyde were heated on the steam bath until the mixture became homogeneous. Work-up aa above gave the product aspaleyellowcrystals(0~55 g,83 %),m.p. 21>215”.(Found: C, 78.80; H, 5.29; N,499. talc. for C,,H,,NOz: C, 78.87 ; H. 5.22 ; N, 4.84 %.) N-Pheny&4-pyridylidenee)succinimide (7~). A slight excess of pyridinetiboxaldehyde was slowly added to a solution of 2.18 g (5 mmole) of lc in 20 ml MeOH. A rapid reaction occurred. Upon cooling, the product crystallized. This was liltered and washed with MeOH giving.l.1 g (83 Pi,). m.p. 210-212”. (Found: C. 72.46; H. 4.58; N. 10.38. Calc. for ClbH1zN202: C 72.71; H. 457; N. 1@60%.) 2i2,2.1-Bicyclohept-5-ene)methylene-N-phenylsuc (76). A mixture consisting of a slight excess of norbornene-2-carboxaldehyde and 2.18 g (5 mmole) of lc was heated on the steam bath for 1 hr. Upon addition of MeOH 10 g (72 %) of the crystalline product was obtained, m.p. 175177”. (Found: C, 77.12; H, 6.16; N. 499. Calc. for C,,H,,NOz: C, 77.39; H, 6.13; N, 5.01x.) N-Phenyl(2-pyrrylmethylene)succinimide (7e). A slight excess of pyrrole-2-carboxaldehyde and 2.18 g (5 nunole) of lc were mixed and heated on the steam bath for 5 min. The mixture was cooled and triturated with cold MeOH. After filtration, the ppt was recrystallized from MeOH giving 1.1 g (87 %), m.p. 24&248”. (Found:C.71.41;H,4.76;N, 11~05.Calc. forC,,H,,N,O,:C,71~45;H,479;N, 11.10x.) Furfuylidene-N-phenylsuccinimide (ll). The reaction between furfural and lc was carried out and workedup as above. A 76% yield was obtained and after recrystallization from MeOH had m.p. 197-199”. (Found: C, 71.76; H.4.35; N, 5.62. talc. for C,,H,,NO,: C, 71.13; H,4.37; N, 5.53:/,.) (2-3,4-Dihydro-2H-pyrony[)methylene-N-ph (7g). The reaction between 3,4dihydro-ZHpyran-2-carboxaldehyde and lc was carried out as above giving a 67 % yield, m.p. 156158”. (Found: C, 71.22; H. 5.67; N 5.41. Calc. for C,,H,,NO,: C, 71.35; H, 5.61; N, 5.20x.) (2-Thiophy&nethylene-N-phykwxinimide (7l1). The reaction between lc and thiophene-2-auboxaldehyde was carried out as above giving an 83 % yield, mp. 249-250”. (Found: C, 6713; H, 417; N, 4.86. Calc. for C,SHIINOzS: C. 689; H, 4.11; N, 5.19x.) Benzylidenesuccinie @a). Triphenylphosphoranylidenesuccinimide (10 g, 2.8 mmoles) and a slight excess of benzaldehyde were mixed. An exothermic reaction occurred. After the initial reaction the mixture was warmed on the steam bath for 3 min and let stand at room temp for 1 hr. Ether was added and crude product filtered off. After recrystallization from MeOH, @4 g (77 %) was obtained, m.p. 198-200”. (Found : C, 7@78; H, 487; N,,7.31. Calc. for C,,H,NO,: C 7@57; H, 4%; N, 7.48x.) (4-Pyridinylidene)succinimide (8b). Triphenylphosphoranylidenesuccinimide (3.6 g 001 mole) and a slight excess of pyridine4carboxaldehyde in 30 ml DMSO were stirred for 2 hr at m”. The mixture was then heated at the b.p for a few min until homogeneous soln was obtained After cooling, crystailization occurred giving 1.75 g (92 %) of a white cry&line product, m.p. 195-197”. (Found: C, 63.31; H, 4.26; N, 1470. Calc. for C,,HsN,O,: C, 63.82; H, 4.28; N. 14.88x.) Propylidenesuccinimide (8e). Maleimide (la g, 001 mole) and tri-n-butylphosphine (20 g, 001 mole) in 15 ml glacial AcOH were warmed on the steam bath for a few min and then let stand for 30 min. The solvent was evaporated at 50” (1 mm) leaving a pinkish oil residue. Distilld propionaldehyde (3 g) was added to the oil and the mixture was warmed on the steam bath and then let stand for 3 hr. Excess aldehyde

2252

E. HEDAYAand S. T~~OROPULOS

was removed in uacw giving an oil. This was chromatographed on Fluorisil packed in hexane. Elution with ether gave a white crystalline material, m.p. 65-70”, in 40% yield. This was purified by sublimation. (Found : C 6008; H, 652; N, 1006 Calc for C,H,NOs : C, @42; H, 652; N, 1006 “/,) N,N-Bibenzyltdenesuccfnimide (9a). Compound 2 (14 g, 1.7 mmoles) and excess benxaldehyde were heated on the steam bath for 10 mitt The mixture was then let stand at room temp for 3-4 hr. The reaction mixture was then triturated with boiling MeOH, cooled, and filtered. The ppt was recrystallixed from benxette giving 06 g (82’4, m.p. 270-272”. (Found: C, 7107; H, 4.13; N, 750. Calc. for C,,HsNOs: C, 7095 ; H. 4.33 ; N. 7.52 %.) N,N’-Bic&tnamaffdettesuccinfmide (9b). N,N’-Bitriphenylphosphoranylidenesuccittimide (@25 g, 034 mmole) and excess cinnamaldehyde were heated on the steam bath for a few min until the mixture became homogeneous. After cooling, ether was added. The ppt was washed thoroughly with more ether giving 0.14 g (cr. 100%). mp. 280-282”. (Found: C, 73.42; H, 4.71; N, 651. Calc. for C,,H,,NO,: C, 73.57; H, 4.74; N, 6.59 %.) N,N’-Motpholino(4-pyridylidcnc)sucn’nimi (10). Compound ld (2.2 g) and a slight excess of pyridine4-carboxaldehyde were combined. After an initial exothermic reaction, the mixture was WaRMd to 40-50 for 50 min. The reaction mixture was then chromatographed on Fluorisil packed in hexane. A So/50 mixture of acetone-ether eluted 0.8 g (59%) of product which was recrystallixed from MeOH, m.p. 227-229”. (Found: C 6195; H, 554; N, 1508. Calc for C,,H,,N,O,: C, 61.52; H, 5.53; N, 15.37x.) 1,4-Bi4N-phenylsuccinimidylidene)butane (11). To a soln of lc (2.17 g, 5 mmole) in 20 ml A&H, 3 ml 2,5-dimethoxytetrahydrofuran was added. The soln was stirred overnight at 50-60”. After cooling and filtration a ppt was obtained (@5 g, 50%) which after thorough washing with acetone had ap. 315-317”. (Found: C. 71.30; H, 508; N, 6.85.Calc for C,,H,,NO,: C 71.98; H, 508; N, 699%) CyclohexyMme-N-phenyhccinimide (12). N-phenylmaleimide (1.75 g, O-01mole) and 2.20 g (Ml mole) tri-n-butylphosphine in 20 ml glacial AcOH were heated on the steam bath for 30 min. The solvent was removed in wctw. The residue was washed with hexane and then combined with excess cyclohexanone. The mixture was heated on the steam bath for 4 hr and the excess cyclohexanone was removed in uucuo. The residue crystallixed upon addition of MeOH-water. Recrystallization from MeOH gave 05 g (2Oy& m.p. 104-106”. (Found: C, 74.87; H, 664; N, 5.62. Calc. for CtdH,,N02: C, 75.26; H, 671; N, 5.48%) Dicfrvuunalidenesuccinic Mhydrfde (IS). A mixture of lr (10 g, 28 mmole), excess cinnamaldehyde, and 30 drops of glacial AcOH were heated at 60”. Within a few min the reaction mixture became homogeneous and then a red ppt formed. The ppt was washed thoroughly with MeOH giving 0.5 g (55 %) of red product, m.p. 236238” (lit.” m.p. 2159. This material could be recrystallixed from acetonitrile. fn some cases small amounts of a more soluble isomeric product having m.p. 18&200” could be. obtained in this way. Tbe IR of this material differed in the fingerprint region from that of the higher melting product. (Found: C, 8@27; H. 488. Calc. for C,,H,,O,: C, 8048; H, 487x.) The same material could bc obtained in about @S% yield from the Stobbe condensation of sodium succinate (005 mole), cinnamaldehyde @075 mole), and Ac,O (@05 mole) in DMF (30 ml). The reaction was carried out at 80” for 5 hr. Di(phenylpropmgylidene)succininc hydride (19). Phenylpropargyl aldehyde (3.0 g) was added to a stirring soln of 3.6 g (001 mole) of la in 50 ml of ice-cold MeOH. After stirring for I hr, the reaction mixture was filtered and the ppt (I.75 g, 55 %) was washed thoroughly with cold MeOH, mp. 133-I 35”. (Found : C. 81.14; H. 3.71. talc. for C,,H,,O,: C. 8146; H. 3.73 ‘4.) The IR had characteristic acetylene and anhydride bands at 4.5. 550 and 568 u. N’-Ani/ino(N”-piperodinyl)methylene-N”’-pknylsuccinimide (k). Phenylisocyanate (1 ml) and 2 17 g (5 mmole) of lc were mixed at room temp. An exothermic reaction occurred. After 15 min the excess isocyanate and the triphenylphosphine oxide was removed by washing the ppt first with hexane and then with benzene. The crude ketenimine could be recrystallized from benzene giving a white product in 79% yield. mp. I561 56”. The IR of this material had characteristic bands at 490, 570.5.85 sh, and 594 p. The NMR in CDCls consisted of bands at 3.80 s and 7.45 m ppm. The ketenimine was stirred with excess piperadine on the steam bath for 5 min. After cooling, the product was precipitated by addition of MeOH. After filtration and washing with acetone I.1 g (61 “/‘was obtained, m.p. 172-174”. (Found: C. 72.71; H, 6.43; N. 11.58. Calc. for CssH,,N,O,: C, 73.10; H, 6.41; N, 11.62%.) N’-Anifiaudfethylmninomet/tyleue-N”-phenylsuccinimfde (21b). The adduct lc (435 g. IO mmole) was converted to the crude ketenimine as above. This was refluxed with excess EtsNH for 15 min. The excess &NH was removed in uacuo and upon addition of MeOH to the residue, a crystalline product was

The preparation and reactions of stable phosphorus ylides

2253

obtained in 96% yield, m.p. 149-151”. (Found: C, 71.97; H, 667; N, 11%. Calc. for C,,H,sNO,: C. 72.18 ; H. 663 ; N. 12.02%.) N’-Anilino(N”-aziridinynmethy~-N”’-p~nyl~cc~~~e (21~). The adduct lc (4.35 g, 10 mmole) was converted to the crude ketenimine as above. This was combined with a&dine at room temp. whereupon an exothermic reaction occur&. After the reaction mixture cooled it was worked up as above giving 2.5 g (78?Q. m.p. 143-145”. (Found: C. 71.26; H, 5.45; N, 1299. Calc. for C~gH,,N,O,: C, 71.45; H, 5.36; N. 13.15%.) Methylamino(N’-1.2.5,Ctetrahydropyriinyl)methylene-N”-phenylsuccinimide (21d). The adduct lc (2.18 g, 5 mmole) was refluxed with excess methyl isocyanate for 10 min. Theexctss isocyanate was removed in vacua and the residue was stirred with excess 1,2,5.6-tetrahydropyridine at 60” for 15 min. After cooling, the product was obtained by addition of MeOH in 70% yield, m.p. 150-152”. (Found: C. 68.52; H. 647; N. 14.19. Calc. for C,,HL9N301: C. 6866; H. 644; N, 14.13%) N’-Anilino(S-phenyllhio)methylene-N”-ph~yl~ccinimide (23). The crude 208 was obtained from 2.18 g adduct Ic and phenyl isocyanate as above. This was mixed with excess benzenethiol and heated for a few min at 50-60”. whereupon an exothermic reaction occurred. The reaction mixture was allowed to cool to room temp. Excess benzenethiol was removed by washing the crude mixture with hexane. The residue was chromatographed on Fluorisil packed in ether. Ether eluted an oil which after washing with hexane was recrystallized from cold ether giving 1.1 g (57%). mp. 109-l 11”. (Found: C, 7150; H. 478; N. 7.19. Calc. for C,sH,,N,O,S: C 71.47; N. 4.69; N. 725x.) 3-(N’-Anilino)N”-phenylmaleimide (24). The adduct lc (2.18 g 5 mmole) and 06 g nitrosobenzene in 45 ml MeOH were stirred at 50” for 15 min. The reaction mixture was cooled and let stand at room temp for 30 min. After liltration. washing with MeOH and recrystallization from acetone I.32 g (ca. 99 “4) was obtained. mp. 238-240” (lit.16 m.p. 2M”). (Found: C. 72.53; H. 4.56; N. 1082. Calc. for C,,H,,N,O,: C. 72.71 ; H. 4.57; N. IO%0?&.) The IR corresponded exactly with that published I6 for 24. The NMR obtained in DMSOd6 had resonances at 5.82 s and 7.4 m ppm in an area ratio of 1 : 10. Acknowledgement-The

authors acknowledge the assistance of Mr. S. Stournas and Miss Mildred Kent on the fulgide synthesis as well as helpful discussions with R. L. Hinman and P. deMayo.

REFERENCES

’ a E. Hedaya. R. L. Hinman and S. Theodoropulos. J. Org. Chem. 31. 13 11 (1966); b Ibid. 1317 (1966). ’ ’ R. F. Hudson and P. A. Chopard, Helu. Chim. Acta 46.2178 (1963); b P. A. Chopard and R. F. Hudson. Z. Noturforsch 18b. 509 (1963); ’ C. Osuch. J. E. Franz and F. B. Zienty, J. Org. Chem 29.3721 (1964); ’ G. Aksnes. Acta. Chem. Stand. 15.692 (1961); ’ A. Sch6nbcrg and A. F. A. Ismail J. Chem Sot. 1374 (1940). ’ W. F. Johnson and G. H. Daub, Organic Reactions 6. 1 (1951). 4 R. J. Cotter. C. K. Sauers and J. M. Whelan. J. Org. Chem 26, 10 (1961). ’ ’ D. Y. Curtin and L. L. Miller. Tetrahedron L.etters1869 (1965); b M. L. Ernst and G. L. Schmir, J. Am. Chem. Sot. 88 5001 (1966); and references cited therein. ’ D. Y. Curtin and L. L. Miller, Ibid. 89.637 (1967). 6 A. W. Johnson. Ylide Chemistry p. 138. Academic Press. New York, N.Y. (1966). ‘I ’ Ch. Rflchardt. A. Eichler and P. Panse. Angew Chem. 75.858 (1963); b S. Fliszar. R. L Hudson and G. Salvadori. Hela. Chim. Acta 47. 159 (1964). ’ W Rudiger and W Klose. Tetrahedron Letters No 47. 5893 (1966). ’ ’ Ref. 6. p. 177; b R. F Hudson. Structure and Mechanism in Orgam+Phaspharus Chemistry p. 229. Academic Press. New York. N.Y. (1965). lo H. 0. House and G. H. Rasmussen, J. Org. Chem 26.4278 (1961). ‘I ’ L. Batt. Liebigs Ann. 331, 160 (1904); b R. Kuhn and A. Winterstein, Helu. Chim. Acta 11 (1928).

2254 l2 I3 I* Is I6 I7

E. HEDAYAand S. THL?O~ROPULOS

H. Staudinger and J. Meyer. H&J. Chim. Acta 2,635 (1919). S. Trippett and D. M Walker, J. Ckem. Sot. 3874 (1959). C. L. Stevens, R. C. Freeman and K. Nell, J. Ore. Chem 30.3718 (1965). A. Schonlnxg and K. H. Brosowski, Chem. Ber. 92 2602 (1959). W. S. Awad. S. M A. R. Omran and F. Nagieb, Tetrahedron 19, 1591 (1963). ’ A. M El-Abbady, S. H. Doss. H. H. Moussa and M Nossier, J. Org. Chem. 26 4871 (1961); b L. S. El-Assal, S. M. A. El-Wahhab and N. S. Baceili. J. Chem. Sot. 740 (1962). ‘s ’ J. Meyer. Helu. Chim Acto 40, 1052 (1957); b G. Luscher. Dissertation, Eidg. Techn. Hochschule, Zurich (1922). ” G. R. Harvey. J. Org. Chem. 31. 1587(1965). 2o a Ref. 6. p. 111; * S. Tripett. Quarterly Rev. 17, 406 (1964).